CN102053983B - Method, system and device for querying vertical search - Google Patents
Method, system and device for querying vertical search Download PDFInfo
- Publication number
- CN102053983B CN102053983B CN 200910210422 CN200910210422A CN102053983B CN 102053983 B CN102053983 B CN 102053983B CN 200910210422 CN200910210422 CN 200910210422 CN 200910210422 A CN200910210422 A CN 200910210422A CN 102053983 B CN102053983 B CN 102053983B
- Authority
- CN
- China
- Prior art keywords
- server
- commodity classification
- query
- commodity
- user
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Landscapes
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The embodiment of the invention discloses a method, a system and a device for querying a vertical search, wherein the method comprises the following steps: a query server acquires the query information of a user; the query server acquires a query result from a lookup dictionary of a commodity category server according to the query information, wherein the query result refers to the commodity sub-categories under the commodity category matched with the query information and corresponding weight of the commodity sub-categories; and the query server sorts the commodity sub-categories in the query result according to the corresponding weight thereof, and sends the obtained sort result to the user, so that the user looks up the commodity sub-categories, and a log server generates a log according to the commodity category looked up by the user and the query information, and sends the log to an analysis server to carry out statistical analysis and then obtain a statistical analysis result, wherein the statistical analysis result is used for updating the lookup dictionary of the commodity category server for subsequent query. Through the method, the system and the device provided by the embodiment of the invention, the query result returned to the user according to the click records of the user is implemented, and the correlation between the query result and the user query is improved.
Description
Technical field
The application relates to networking technology area, particularly relates to a kind of querying method, system and device of vertical search.
Background technology
Growing along with the internet, the canned data amount is huge day by day on the internet.When people need obtain the specific information of certain aspect, search for by search engine.But because the quantity of information on the internet is excessive, the Search Results that adopts the universal search mode to obtain lacks accuracy, so the vertical search mode has obtained development fast.Vertical search is the professional search engine at some industries, be segmentation and the extension of search engine, be that the special information of certain class in the web page library is once integrated, directed branch field extracts and returns to the user with certain form again after the data that need are handled.Relatively universal search engine contain much information, inquire about new search engine service pattern inaccurate, that the degree of depth is not enough etc. puts forward, by the information that certain value is arranged and the related service that provides at a certain specific area, a certain specific crowd or a certain particular demands.Its characteristics are exactly " special, smart, dark ", and have the industry color, the magnanimity information disordering of the universal search engine of comparing, and vertical search engine then seems absorbed more, concrete and gos deep into.
The application direction of vertical search engine is a lot, such as enterprise's library searching, supply-demand information search engine, shopping search, house property search, talent's search, map search, mp3 search, picture searching etc., almost all trades and professions various information can further be refined into all kinds of vertical search engines.
When vertical search is used for the shopping search; the user is at B2C (Business to Customer; business to consumer's shopping mode) or C2C (Consumer to Customer; consumer to consumer's shopping mode) shopping website input inquiry word shopping; usually can return two-part result: 1. the navigation information of commodity classification, 2. Search Results associated with the query.The commodity classification title of navigation is got up according to the structure organization of tree, make things convenient for the user along the path of tree construction from top to bottom the information by commodity classification navigate to Search Results more accurately.
Commodity classification tree construction is kept at the corresponding tables of data of database, and the input of data need manually be carried out with maintenance, and the displaying of each commodity all must belong to some nodes or a plurality of node of this commodity classification tree in B2C or C2C website.
Current e-commerce website often commodity amount is too huge, causes commodity classification too much.On the commodity amount of more than one hundred million scales, commodity classification tree usually can be near 10,000 nodes, and the classification number of nodes of each level tends to nearly tens.When the user inquired about, the Taxonomy Information that is shown to the user was too much, and can't tell these commodity classifications of user which is more important to user's inquiry.To this problem, the settling mode of main flow is when the user inquires about at present, adds up each class return results quantity now one by one.Then these commodity classifications according to commodity amount according to sorting from big to small, and certain threshold values is set.The classification that the commodity number is lower than this threshold values stashes.Reach the purpose that reduces classification quantity.
In the process that realizes the application, the inventor finds prior art, and there are the following problems at least:
(1) classification of Xian Shiing and user's inquiry correlativity is very low.
(2) there is not mechanism to determine which commodity classification is more important between the commodity classification.
(3) quantity that shows for the classification of commodity only stashes the high classification of correlativity with threshold values control meeting.
Summary of the invention
The embodiment of the present application provides a kind of querying method, system and device of vertical search, is used for improving the correlativity of Query Result and user's inquiry.
The embodiment of the present application provides a kind of querying method of vertical search, is applied to comprise in the system of querying server, Analysis server and log server, it is characterized in that, comprising:
Querying server obtains user's Query Information;
Described querying server obtains Query Result according to described Query Information in the queries dictionary of described commodity classification server, described Query Result for the commodity class that is complementary with described Query Information now sub-commodity classification and corresponding weight;
Described querying server sorts the sub-commodity classification in the described Query Result according to its corresponding weight, and ranking results sent to described user, described user is checked, and make log server generate daily record according to commodity classification and the described Query Information that described user checks, and described daily record is sent to Analysis server carry out statistical study and obtain statistic analysis result, described statistic analysis result is used for upgrading the queries dictionary of described commodity classification server, is used for follow-up inquiry.
Wherein, at described querying server before corresponding weight sorts according to it with the sub-commodity classification in the described Query Result, also comprise: set a weight threshold, the sub-commodity classification of the weight in the described Query Result greater than described weight threshold sorted according to its corresponding weight.
Wherein, described querying server obtains before user's the Query Information, also comprises:
Front-end server obtains user's Query Information, and described Query Information comprises described user's query word and commodity classification;
Described front-end server carries out normalized to described query word and obtains the commodity ID of commodity classification correspondence;
Described front-end server will be transmitted to described commodity classification server through described query word and the described commodity classification ID of normalized.
Wherein, it is characterized in that,
The commodity classification that described log server is checked according to described user and described Query Information generate before the daily record, also comprise:
Obtain the described user's of front-end server forwarding Query Information;
Analysis server carries out statistical study and obtains statistic analysis result, and described statistic analysis result is used for upgrading the queries dictionary of described commodity classification server, is used for follow-up inquiry and specifically comprises:
Receive the interior daily record of Preset Time of described log server timed sending;
Carry out statistical study according to the described daily record in the Preset Time, obtain statistic analysis result, described statistic analysis result is commodity classification and the corresponding weight that described user checks; Described weight comprises the number of clicks of the commodity classification correspondence that described user checks and corresponding click probability in commodity classification at the same level;
According to commodity classification tree, with described statistic analysis result generated query file;
Described inquiry file is sent to described querying server, so that described querying server upgrades the queries dictionary of described commodity classification server, the inquiry that the user is follow-up according to described inquiry file.
Wherein, described querying server sorts the sub-commodity classification in the described Query Result according to its corresponding weight, and ranking results is sent to before the described user, also comprises:
Described querying server splices described Query Result, and described splicing comprises the commodity classification that obtains the ID of commodity classification described in described Query Result correspondence.
The embodiment of the present application provides a kind of inquiry system of vertical search, it is characterized in that, comprising:
Querying server is for the Query Information that obtains the user; Obtain Query Result according to described Query Information in the queries dictionary of described commodity classification server, described Query Result for the commodity class that is complementary with described Query Information now sub-commodity classification and corresponding weight; Sub-commodity classification in the described Query Result is sorted according to its corresponding weight, and ranking results is sent to described user, described user is checked; Obtain the statistic analysis result that Analysis server sends, upgrade the queries dictionary of described commodity classification server according to described statistic analysis result, be used for follow-up inquiry;
Log server, the commodity classification and the described Query Information that are used for checking according to described user generate daily record, and described daily record is sent to described Analysis server.
Analysis server is used for receiving the described daily record that described log server sends; Statistical study is carried out in described daily record obtained described statistic analysis result; Described statistical study binded up one's hair give described querying server.
Wherein, described Query Information comprises query word and commodity classification, and described log server comprises:
Acquisition module is for the user's who obtains the front-end server forwarding Query Information;
Generation module, the commodity classification and the described Query Information that are used for checking according to described user generate daily record;
Sending module is used for that the described daily record that described generation module generates is sent to described Analysis server and carries out statistical study acquisition statistic analysis result, and described statistic analysis result is used for upgrading the queries dictionary of described commodity classification server, is used for follow-up inquiry.
Wherein, described Analysis server comprises:
Receiver module is used for receiving the daily record that described log server sends;
Statistical analysis module is used for statistical study is carried out in the described daily record that described receiver module receives, and obtains statistic analysis result;
Sending module, the described statistic analysis result that is used for described statistical analysis module is obtained sends to querying server, makes described querying server upgrade the queries dictionary of described querying server, is used for follow-up inquiry.
Wherein, described statistical analysis module comprises:
The statistical study submodule is used for carrying out statistical study according to the described daily record that the described acquisition module in the Preset Time obtains, and obtains statistic analysis result, and described statistic analysis result is commodity classification and the corresponding weight that described user checks; Described weight comprises the number of clicks of the commodity classification correspondence that described user checks and corresponding click probability in commodity classification at the same level;
Generate submodule: be used for according to commodity classification tree the described statistic analysis result generated query file that described statistical study submodule is obtained.
The embodiment of the present application provides a kind of server, as querying server, is applied to comprise in the system of querying server, Analysis server and log server, it is characterized in that, comprising:
Acquisition module is for the Query Information that obtains the user;
Enquiry module, the described Query Information that is used for obtaining according to described acquisition module obtains Query Result in the queries dictionary of described commodity classification server, described Query Result for the commodity class that is complementary with described Query Information now sub-commodity classification and corresponding weight;
Sending module, sub-commodity classification for the described Query Result that described enquiry module is obtained sorts according to its corresponding weight, and ranking results sent to described user, described user is checked, and make log server generate daily record according to commodity classification and the described Query Information that described user checks, and described daily record is sent to Analysis server carry out statistical study and obtain statistic analysis result, described statistic analysis result is used for upgrading the queries dictionary of described commodity classification server, is used for follow-up inquiry;
Update module is used for upgrading according to the statistic analysis result that described acquisition module obtains the queries dictionary of described commodity classification server, and the queries dictionary after the described renewal is sent to described enquiry module, is used for follow-up inquiry.
Wherein, described sending module also is used for: set a weight threshold, the sub-commodity classification of the weight in the described Query Result greater than described weight threshold sorted according to its corresponding weight.
Wherein, described Query Information comprises through the described query word of normalized and the commodity classification ID of described commodity classification correspondence;
Also comprise concatenation module, be used for described Query Result is spliced that described splicing comprises the commodity classification that obtains the ID of commodity classification described in described Query Result correspondence.
The application has improved the correlativity of Query Result and user inquiry by click the Query Result that record returns the user according to the user.Certainly, arbitrary product of enforcement the application might not need to reach simultaneously above-described all advantages.
Description of drawings
In order to be illustrated more clearly in the application or technical scheme of the prior art, to do simple the introduction to the accompanying drawing of required use in the application or the description of the Prior Art below, apparently, accompanying drawing in describing below only is some embodiment of the application, for those of ordinary skills, under the prerequisite of not paying creative work, can also obtain other accompanying drawing according to these accompanying drawings.
Fig. 1 is the process flow diagram of a kind of querying method of vertical search in the embodiment of the present application;
Fig. 2 is the interaction figure of a kind of querying method of vertical search in the embodiment of the present application;
Fig. 3 is the interaction figure of a kind of querying method of vertical search in the embodiment of the present application;
Fig. 4 is the process flow diagram of a kind of querying method of vertical search in the embodiment of the present application;
The click classification tree that Fig. 5 generates for the commodity classification of checking according to click in the embodiment of the present application;
The click classification tree that Fig. 6 generates for the commodity classification checked according to click in the embodiment of the present application and number of times;
Fig. 7 is the process flow diagram of a kind of querying method of vertical search in the embodiment of the present application;
Fig. 8 is the process flow diagram of a kind of querying method of vertical search in the embodiment of the present application;
Fig. 9 is the structural representation of a kind of log server in the embodiment of the present application;
Figure 10 is the structural representation of a kind of Analysis server in the embodiment of the present application;
Figure 11 is the structural representation of a kind of Analysis server in the embodiment of the present application;
Figure 12 is the structural representation of a kind of querying server in the embodiment of the present application;
Figure 13 is the structural representation of a kind of querying server in the embodiment of the present application.
Embodiment
The embodiment of the present application proposes: querying server obtains user's Query Information; Described querying server obtains Query Result according to described Query Information in the queries dictionary of described commodity classification server, described Query Result for the commodity class that is complementary with described Query Information now sub-commodity classification and corresponding weight; Described querying server sorts the sub-commodity classification in the described Query Result according to its corresponding weight, and ranking results sent to described user, described user is checked, and make log server generate daily record according to commodity classification and the described Query Information that described user checks, and described daily record is sent to Analysis server carry out statistical study and obtain statistic analysis result, described statistic analysis result is used for upgrading the queries dictionary of described commodity classification server, is used for follow-up inquiry.
Below in conjunction with the accompanying drawing among the application, the technical scheme among the application is carried out clear, complete description, obviously, described embodiment is a part of embodiment of the application, rather than whole embodiment.Based on the embodiment among the application, the every other embodiment that those of ordinary skills obtain under the prerequisite of not making creative work belongs to the scope that the application protects.
As stated in the Background Art, the application direction of vertical search engine is a lot, such as enterprise's library searching, supply-demand information search engine, shopping search, house property search, talent's search, map search, mp3 search, picture searching.Give an example to illustrate and to be more readily understood, such as shopping search engine, overall flow is roughly as follows: according to user's searching requirement, after grasping webpage, webpage merchandise news is extracted, extract trade name, price, brief introduction ... even can further the notebook brief introduction be subdivided into " brand, model, CPU, internal memory, hard disk, display screen ... ", and Search Results returned to the user.For the searching requirement according to the user, improve the correlativity of return message and the information that is used for searching for, the application has proposed a kind of querying method of vertical search.
The embodiment of the present application provides a kind of querying method of vertical search, as shown in Figure 1, may further comprise the steps:
Wherein, described Query Information can comprise the query word of described user's input and the commodity classification of user's input or selection.
Wherein, at described querying server before corresponding weight sorts according to it with the sub-commodity classification in the described Query Result, can also set a weight threshold, the sub-commodity classification of the weight in the described Query Result greater than described weight threshold sorted according to its corresponding weight.By the Query Information commodity classification that the inquiry acquisition is complementary in the inquiry dictionary, can sort according to weight to the sub-commodity classification of now all of this commodity class, also can only sort according to weight to parton commodity classification.When parton commodity classification is sorted according to weight, can preestablish a weight threshold, the weight of only choosing in the Query Result sorts according to its corresponding weight greater than the sub-commodity classification of this weight threshold.By parton commodity classification is sorted, and ranking results is sent to the calculated amount that the user can reduce ordering, improve the computing velocity of ordering.
The querying method of a kind of vertical search that provides in the embodiment of the present application, as shown in Figure 2, wherein, querying server is specially commodity classification querying server, Analysis server is specially distributed document storage and concurrent operation platform, as shown in Figure 3, specifically may further comprise the steps:
(1) front-end server receives the Query Information of user's input, comprises query word and commodity classification in the Query Information, and front-end server is transmitted to the log query server with this query word, is used for follow-up according to this time inquiry generation daily record; This Query Information is transmitted to commodity classification querying server, is used for inquiring about according to this Query Information at commodity classification querying server, and Query Result is returned to the user, click for the user and check.
(2) user clicks commodity classification and commodity according to the needs of oneself at front-end server and checks, at this moment, each click checks that action all can generate a corresponding daily record record by log server, the commodity that this daily record has the user to click to check and the commodity classification under this commodity.After a period of time, log server imports to distributed document storage and concurrent operation platform with all daily records in this section period, is used for storage and statistical study to daily record.
(3) the click analytic statistics program in distributed document storage and the concurrent operation platform is carried out statistical study to the daily record in a period of time, this analysis comprises obtains the weight that each commodity classification is clicked, weight has embodied the degree of correlation of commodity classification and user's Query Information, preferably, can be number of clicks or the click probability of Query Information correspondence.When commodity are checked in user's click, because each daily record all records this time click and checks corresponding user's query word, and commodity classification under the commodity checked of this click, thereby just can obtain at same query word commodity classification and the weight at the commodity place that the user checks according to a large amount of daily records.Therefrom can know, when the user imports this query word, the degree of correlation of each commodity classification and this query word, the i.e. degree of user's care.Click analytic statistics program as output, is sent to commodity classification querying server with statistic analysis result.
(4) commodity classification querying server passes through the hash algorithm with the form compiled query dictionary of Key-Value according to statistic analysis result, thereby improves the speed of inquiry.Wherein Key is the query word of user's input, and Value is a plurality of commodity classifications and the corresponding weight of this query word correspondence, and this queries dictionary has embodied the degree of correlation of query word and each commodity classification.When the query word of knowing the user and commodity classification, namely can in queries dictionary, inquire about, commodity classification that will be relevant with this query word returns to the user, and according to the height of the degree of correlation, namely the weight of commodity classification is arranged, and selects for the user.
Need explanation the time, upgrading queries dictionary according to this statistic analysis result can be the queries dictionary that only generates according to the Query Result of this time statistical study, also this time statistic analysis result can be added in the former queries dictionary, generate new queries dictionary, also can delete in the set period as required or set period before statistic analysis result corresponding data in queries dictionary.For example: for the commodity of clothes class, when to season of changing according to the season, statistic analysis result before will not be suitable for inquiry subsequently, thus will be not used in the data deletion of the statistic analysis result correspondence of inquiry subsequently, to guarantee the accuracy of inquiry.
By above-mentioned explanation as can be known, this querying method is the process of a circulation.When the user inquired about, to be commodity classification querying server checked inquiry in the queries dictionary that behavior generates and the Query Result that obtains in the click of a last time period to the Query Result that returns; To carry out statistical study in the next time period and in this inquiry, click the daily record that the behavior checked generates according to the user, and send to commodity classification querying server, be used for follow-up commodity classification inquiry.
The embodiment of the present application provides a kind of querying method of vertical search, as shown in Figure 4, may further comprise the steps:
Herein, it is query word that our definition is used for carrying out the required Query Information of one query, also can comprise the commodity classification.
Front-end server provides with the user and carries out mutual window by query page.The user visits this query page by browser, and the Query Information that input is wanted to inquire about in this query page is inquired about.
For example, when the user need buy a T-shirt, the user logined the online transaction website, and by the query page of this website, the input user wants the Query Information " T-shirt " inquired about.At this moment, when if the user only wants to check the information of relevant men's clothing T-shirt, the combobox that system ejects in the time of can passing through input " T-shirt " is selected the commodity classification, thereby in the input inquiry word, input commodity classification limits query context, and for example: the user selects " the T-shirt men's clothing " in the combobox, wherein " men's clothing " is the commodity classification, and Query Information is " T-shirt men's clothing ".
The user is input inquiry word " T-shirt " only, when providing " men's clothing ", " women's dress " and " easy dress " on the page that system returns according to " T-shirt ", the user selects " men's clothing ", and wherein " men's clothing " is the commodity classification, and Query Information is " T-shirt men's clothing ".
Front-end server gets access to the Query Information of user's input by query page, specifically may further comprise the steps:
(1) front-end server obtains the Query Information of user's input by query page.
(2) front-end server obtains the corresponding commodity classification of the commodity classification ID in this Query Information.
Inquiry for the ease of the commodity classification querying server of rear end, front-end server is not directly the commodity classification in the Query Information that obtains to be sent to commodity classification querying server, but the commodity classification ID of this commodity classification correspondence is sent to commodity classification querying server, so front-end server need obtain the commodity classification ID of this commodity classification correspondence.
At this moment, if Query Information is query word only, do not have the commodity classification, then need not to obtain commodity classification ID.
Certainly, also can the commodity classification be transmitted to the commodity classification querying server of rear end by front-end server, obtain the commodity classification ID of this commodity classification correspondence by the commodity classification querying server of rear end, inquire about by this commodity classification ID again.
The commodity classification querying server of front-end server and rear end carries out the transmission of data by interface.The input and output of front-end interface adopt the mode of http to conduct interviews and return results.Input parameter adopts the http agreement, submits to parameter to give " commodity classification querying server " by the Get mode.The parameter of importing into mainly contains two: query word and commodity classification ID.
The form that input connects is as follows:
Http:// host? query=Cha Xunci ﹠amp; Catid=commodity classification id
Return results adopts the form of XML, and concrete form is as follows:
<？xml?version＝″1.0″encoding＝″GBK″？>
<conf>
<module?name＝″catsort″v＝″″>
<p?name＝″result″>510001:200，510021:100，10221:30</p>
</module>
</conf>
The commodity classification querying server of step 404, rear end is inquired about, and Query Result is sent to front-end server according to the Query Information that receives.
The Query Result that step 405, front-end server will be sent by the commodity classification querying server of rear end is shown to the user.
Wherein, when Query Information only was query word, Query Result was and all commodity classifications of this query word coupling and corresponding weight (being number of clicks or the click probability of commodity classification) thereof, and arranges according to weight order from high to low; When Query Information is query word and commodity classification, Query Result is and the sub-commodity classification of now all of commodity class of this Query Information coupling and corresponding weight (being number of clicks or the click probability of commodity classification) thereof, and arranges according to weight order from high to low.
Preferably, be shown in user's the Query Result page, both comprised the sub-commodity classification of now all of commodity class that is complementary with Query Information, also comprised the detailed list of now each of these sub-commodity classes merchandise news, so that the user directly selects concrete commodity to check in this Query Result.
Front-end server is by the Query Result display page, and the Query Result that the commodity classification querying server of rear end is returned is shown to the user.User's commodity classification that selection will be checked from the Query Result that shows is clicked and is checked, clicks in the subcategory of the commodity classification of checking again, returns to the user for checking by commodity classification querying server according to above-mentioned querying method again.And so forth, find the commodity that to check by checking the commodity classification step by step, click and check.
All added the connection that is saved in " log server " in above-mentioned each shown commodity.In the process that click is checked, after commodity are clicked, all will generate corresponding daily record according to this time click behavior, be kept in the log server.Wherein, click behavior each time is a click logs, and the form of daily record is as shown in table 1:
Table 1 journal format
| Query word | The inquiry classification | Commodity ID | The commodity classification | Item property |
Wherein, query word is the query word in the Query Information of user's input, the inquiry classification is selected commodity to click for user's commodity classification that selected click is checked in the commodity classification that the page returns and from click the page that this commodity classification returns to check, this moment, this commodity classification was the inquiry classification, so it is a plurality of that the inquiry classification may have, and the inquiry classification that the inquiry classification in the daily record is only stored is checked the commodity the last commodity classification of clicking before for the user clicks.After clicking this inquiry classification, the Query Result that returns to the user had both comprised that the sub-commodity classification of now all of this inquiry class also comprised the detailed list of now each of these sub-commodity classes merchandise news.Commodity ID is the corresponding ID of each commodity number, is used for each commodity of unique identification.Commodity classification under the commodity that the commodity classification is checked for the user clicks are direct.This commodity classification may be the sub-commodity classification of inquiry classification.Item property is the corresponding satellite informations of these commodity, for example: brand.
For example, when the query word of user's input is " T-shirt ", in the Query Result that the commodity classification that clicks according to the user returns for " long sleeves T-shirt ", having selected a brand in the detailed list of commodity is the T-shirt of POLO, the commodity ID of this part commodity correspondence is 12200021, its directly affiliated commodity classification is the men's clothing T-shirt, and wherein " men's clothing T-shirt " is the subcategory of " long sleeves T-shirt ", generates a daily record record as shown in table 2 according to above-mentioned information.
Table 2 daily record record
| Query word | The inquiry classification | Commodity ID | The commodity classification | Item property |
| T-shirt | The long sleeves T-shirt | 12200021 | The men's clothing T-shirt | Brand: POLO |
Distributed document storage and concurrent operation platform are used for the daily record that storage generates, and the calculating when all daily records are analyzed.
The generation of each bar daily record record is along with user's click checks that log server regularly should all daily records in the time period import to distributed document storage and concurrent operation platform and produced simultaneously.Wherein, the regular operation of log server can be every day, or per 12 hours etc.
Log server is clicked the analytic statistics program and will regularly be carried out statistical study to stored log after regularly the daily record that generates being imported to distributed document storage and concurrent operation platform, draws statistic analysis result.The object of this statistical study can be nearest ten days daily record, or the daily record in nearest two weeks, and the concrete time can be adjusted according to empirical value or statistical demand.
Concrete, click the analytic statistics program timing stored log is carried out statistical study, draw statistic analysis result, specifically may further comprise the steps:
(1) obtains the daily record of carrying out statistical study.
Log server can regularly upgrade the daily record in distributed document storage and the concurrent operation platform, when click analytic statistics program timing is carried out statistical study to stored log, need obtain the daily record of renewal, to obtain up-to-date daily record, improves statistical accuracy.
(2) query word is carried out normalized.
Because the query word of each user's input not necessarily meets the statistical standard of clicking the analytic statistics program, for the ease of adding up according to query word, need carry out normalized to query word.Normalized comprises removes unnecessary word in the query word, and the conversion of alphabet size between writing carried out, the conversion between the full-shape half-angle, the conversion between the simplified traditional font, the conversion of punctuate and the conversion between the Chinese figure etc. in unnecessary space.Query word through normalized can be directly used in the statistics of clicking the analytic statistics program.
(3) data are carried out in daily record and gather, generate and click distributed data.
When thousands of user inquires about, can import identical query word and click identical inquiry classification, according to the daily record in a period of time, query word and inquiry classification that the user is inquired about gather, and obtain the number of times of the commodity classification of clicking according to each query word.
For example, click the commodity of checking by query word " T-shirt " and have 400, wherein have 200 commodity to belong to men's clothing commodity classification, have 100 commodity to belong to women's dress commodity classification, have 100 commodity to belong to CRUX commodity classification.In belonging to 200 commodity of men's clothing commodity classification, there are 200 commodity to belong to cotta T-shirt commodity classification, there is 0 commodity to belong to long sleeves T-shirt commodity classification.In belonging to 100 commodity of women's dress commodity classification, there are 100 commodity to belong to cotta T-shirt commodity classification, there is 0 commodity to belong to long sleeves T-shirt commodity classification.In belonging to 100 commodity of CRUX commodity classification, there are 60 commodity to belong to lovers and adorn the commodity classification, there are 40 commodity to belong to motion T-shirt commodity classification.
Above-mentioned each click view procedure and all can generate corresponding daily record, according to the gathering of daily record, draw the click distributed data as table 3:
Table 3 is clicked distributed data
| Query word | Click classification: number of times | Click classification: number of times | Click classification: number of times | Click classification: number of times |
| T-shirt | Men's clothing T-shirt: 200 | Women's dress T-shirt: 100 | Lovers' dress: 60 | Motion T-shirt: 40 |
(4) obtain commodity classification tree.
For the ease of user's inquiry and the management of system, all commodity are all classified to it according to attribute, and each commodity has the commodity classification under it, and all commodity classifications have been generated commodity classification tree according to logical order.
Click the analytic statistics program and want to generate the click tree according to the click distributed data, must know commodity classification residing position in all commodity classifications that the user clicks, namely set so need obtain the commodity classification position in commodity classification tree.
(5) in conjunction with commodity classification tree, generate click classification tree according to clicking distributed data.
Comprise all commodity classifications in the commodity classification tree, and represented relation between each commodity classification visually with the form of tree.Click the commodity classification that distributed data has comprised that all users click, and embodied the number of clicks of each commodity classification with the form of literal.Clicking the classification tree namely is that the information in commodity classification tree and the click distributed data is combined, setting with the commodity classification is the form of expression, all add relevant position in the commodity classification tree to clicking each bar data in the distributed data, the number of clicks of the logical relation between the commodity classification of commodity classification tree embodiment and the embodiment of click distributed data is associated, jointly embodies.
At first, generate click classification tree.Construct according to commodity classification and this commodity classification correspondence position in the commodity classification is set clicked in the distributed data, generate and click the classification tree, set as shown in Figure 5 according to the click classification that the click distributed data in the table 3 generates.
Secondly, add the number of clicks of clicking commodity classification in the classification tree.Add the corresponding commodity class of click classification tree to now with clicking the number of clicks of analyzing each commodity classification correspondence in the data, finish the generation of clicking the classification tree, set as shown in Figure 6 according to the click classification that comprises number of clicks that the click distributed data in the table 3 generates.
(6) generate the inquiry file corresponding with clicking the classification tree.
The form of clicking the classification tree has embodied user's click information simply, inquiry for the ease of the commodity classification querying server of rear end, the user's click information that comprises in the click classification need being set is with the form performance of text, so according to clicking the classification number, generate corresponding with it inquiry file, be used for the more queries dictionary of new commodity classification querying server.
Need to prove that for clicking the classification tree, when the commodity classification in clicking the classification tree was not checked by user's click, the number of clicks of this commodity classification in clicking the classification tree was zero degree; For inquiry file, when the commodity classification in clicking the classification tree is not checked by user's click, to can not generate the inquiry file corresponding with this commodity classification number, namely include only in the inquiry file and click the data message of being clicked the commodity classification correspondence of inquiring about in the classification tree by the user.
According to the click classification tree that comprises number of clicks shown in Figure 6, the inquiry file of generation is:
T-shirt root classification men's clothing: 200 women's dresses: 100 CRUXs: 100
T-shirt men's clothing cotta T-shirt: 200
T-shirt women's dress cotta T-shirt: 100
T-shirt CRUX lovers dress: 60 motion T-shirts: 40
To step 408 (6), finished the statistical study to daily record by above-mentioned steps 408 (1), drawn statistic analysis result, be i.e. the number of clicks of the logical relation between the commodity classification of the commodity classification of Dian Jiing, click and each commodity classification.
Need to prove that the number of clicks of commodity classification has embodied this commodity classification at commodity at the same level, i.e. weight in parent all subcategories now under this commodity classification.In addition, this commodity classification weight also can embody with the form of clicking probability in parent all subcategories now under this commodity classification, and this clicks probability by the number of clicks acquisition of correspondence.The weight of above-mentioned inquiry file is to click the form body now of probability, and this inquiry file is:
T-shirt root classification men's clothing: 50% women's dress: 25% CRUX: 25%
T-shirt men's clothing cotta T-shirt: 100%
T-shirt women's dress cotta T-shirt: 100%
T-shirt CRUX lovers dress: 60% motion T-shirt: 40%
The embodiment of the present application provides a kind of querying method of vertical search, as shown in Figure 7, may further comprise the steps:
The commodity classification querying server of step 701, rear end is compiled into queries dictionary with inquiry file.
The commodity classification querying server of rear end need be inquired about the commodity classification according to Query Information when the Query Information that the receiving front-end server sends, and qualified Query Result is sent to front-end server.So when the commodity classification server of rear end was inquired about, the commodity classification that at first needs to be used for inquiry was compiled into queries dictionary, so that the commodity classification querying server of rear end is inquired about.
Queries dictionary is obtained by the compiling of queries dictionary program compiler, particularly, the output result (being inquiry file) that queries dictionary program compiler compiling will be clicked the analytic statistics program is compiled into corresponding Memory Mapping File and its by the hash algorithm with the form of Key-Value, for follow-up inquiry.Wherein Key is the query word of user's input, and Value is a plurality of commodity classifications and the corresponding weight of this query word correspondence.
For example: as follows for the inquiry file that step 408 (6) generates:
T-shirt root classification men's clothing: 50% women's dress: 25% CRUX: 25%
T-shirt men's clothing cotta T-shirt: 100%
T-shirt women's dress cotta T-shirt: 100%
T-shirt CRUX lovers dress: 60% motion T-shirt: 40%
Wherein, by the hash algorithm with the queries dictionary that the form of Key-Value is compiled into to be:
Key:T sympathizes root classification Value: men's clothing 50%; Women's dress 25%; CRUX 25%;
Key:T sympathizes men's clothing Value: cotta T-shirt 100%;
Key:T sympathizes women's dress Value: cotta T-shirt 100%;
Key:T sympathizes CRUX Value: lovers adorn 60%; Motion T-shirt 40%.
The queries dictionary program compiler is carried out the queries dictionary compiling to the output result queries file of clicking the analytic statistics program, is compiled into hash algorithm Memory Mapping File and its with the form of Key-Value, is used for follow-up inquiry.Need to prove that this hash algorithm Memory Mapping File and its can directly be loaded into internal memory to improve system initialization efficient when program initialization.In addition, for convenience of explanation, commodity classification in the embodiment of the present application all adopts the form of expression of commodity classification title, but in practical operation, the form of expression of commodity classification can be the commodity classification ID of title or this commodity classification correspondence of commodity classification, and wherein the form of commodity classification ID is convenient to the commodity classification is inquired about.
The commodity classification querying server of step 702, rear end regularly loads queries dictionary.
Queries dictionary loads and to refer to when startup of server, and the mode by memory-mapped directly is mapped to internal memory to the queries dictionary file with the compiling of hash algorithm Memory Mapping File and its form.
Front-end server provides with the user and carries out mutual window by query page.The user visits this query page by browser, and the Query Information that input is wanted to inquire about in this query page is inquired about, and this Query Information is query word, also can comprise the commodity classification.
For example, when the user need buy a T-shirt, the user logined the online transaction website, and by the query page of this website, the input user wants the Query Information " T-shirt " inquired about.At this moment, when only wanting to check the information of relevant men's clothing T-shirt as if the user, when can pass through the input inquiry word, input commodity classification limits query context, and namely in query page input " T-shirt men's clothing ", wherein " men's clothing " is the commodity classification.
Front-end server gets access to the Query Information of user's input by query page, and is transmitted to the commodity classification querying server of rear end, specifically may further comprise the steps:
(1) front-end server obtains the Query Information of user's input by query page.
(2) front-end server obtains the corresponding commodity classification of the commodity classification ID in this Query Information.
Inquiry for the ease of the commodity classification querying server of rear end, front-end server is not directly the commodity classification in the Query Information that obtains to be sent to commodity classification querying server, but the commodity classification ID of this commodity classification correspondence is sent to commodity classification querying server, so front-end server need obtain the commodity classification ID of this commodity classification correspondence.
Need to prove if Query Information is query word only, do not have the commodity classification, then need not to obtain commodity classification ID.
Certainly, also can the commodity classification be transmitted to the commodity classification querying server of rear end by front-end server, obtain the commodity classification ID of this commodity classification correspondence by the commodity classification querying server of rear end, inquire about by this commodity classification ID again.
The commodity classification querying server of front-end server and rear end carries out the transmission of data by interface.The input and output of front-end interface adopt the mode of http to conduct interviews and return results.Input parameter adopts the http agreement, submits to parameter to give " commodity classification querying server " by the Get mode.The parameter of importing into mainly contains two: query word and commodity classification ID.
The form that input connects is as follows:
Http:// host? query=Cha Xunci ﹠amp; Catid=commodity classification id
Return results adopts the form of XML, and concrete form is as follows:
<？xml?version＝″1.0″encoding＝″GBK″？>
<conf>
<module?name＝″catsort″v＝″″>
<p?name＝″result″>510001:500，510021:100，10221:30</p>
</module>
</conf>
The commodity classification querying server of step 706, rear end obtains Query Information, and the query word in the Query Information is carried out normalized.
Because the query word of each user's input not necessarily meets the query criteria of the commodity classification querying server of rear end, for the ease of inquiring about according to query word, need carry out normalized to query word.Normalized comprises removes unnecessary word in the query word, and the conversion of alphabet size between writing carried out, the conversion between the full-shape half-angle, the conversion between the simplified traditional font, the conversion of punctuate and the conversion between the Chinese figure etc. in unnecessary space.Can be directly used in the inquiry of the commodity classification querying server of rear end through the query word of normalized.
The commodity classification querying server of step 707, rear end is inquired about in queries dictionary according to Query Information, obtains Query Result.
The commodity classification querying server of rear end is inquired about in the queries dictionary that is loaded by step 702 according to Query Information.Because the mode by memory-mapped directly is mapped to internal memory to the queries dictionary file with the compiling of hash algorithm Memory Mapping File and its form, so inquire about by using Hash to search at internal memory in the whole query script, guarantees the efficient of inquiring about.
For example: when the query word of user's input was " T-shirt ", the commodity classification querying server of rear end was inquired about according to the queries dictionary that generates in the step 701, and this moment, Key was " T-shirt ", obtained corresponding Value, and namely men's clothing 50%; Women's dress 25%; CRUX 25%.When the query word of user input be " T-shirt CRUX " or when clicking commodity classification " operation is lain fallow ", the commodity classification querying server of rear end is inquired about according to the queries dictionary that generates in the step 701, this moment, Key was " T-shirt CRUX ", obtain corresponding Value, namely lovers adorn 60%; Motion T-shirt 40%.
By the way, the commodity classification querying server of rear end obtains the Query Result of this Query Information, comprise the commodity classification relevant with Query Information and the corresponding weight of each commodity classification in this Query Result, and arrange according to weight order from high to low.
The commodity classification querying server of step 708, rear end splices the commodity classification information that inquires, and returns to front-end server.
The commodity classification querying server of rear end when inquiring about in queries dictionary according to Query Information, use be query word and the commodity classification ID that in step 705, obtains.And when the commodity classification querying server of rear end sends to front-end server with Query Result, if the commodity classification ID that directly will inquire in queries dictionary sends to front-end server, represent to the user by front-end server, then the user can't know the commodity classification of shown commodity classification ID correspondence, checks thereby can't click.Therefore, the commodity classification querying server of rear end need add the commodity classification information of the commodity classification ID correspondence that inquires in the Query Result to, or after replacing corresponding commodity classification ID, the Query Result that will comprise commodity classification information sends to front-end server, and the user is inquired about by front-end server.
Need to prove, the commodity classification information of the commodity classification ID correspondence that inquires is added in the Query Result, or the operation of replacing corresponding commodity classification ID also can be finished by front-end server.
Because log server regularly imports to the daily record that generates distributed document storage and concurrent operation platform, clicks the analytic statistics program timing stored log is carried out statistical study, the commodity classification querying server of rear end regularly loads queries dictionary.Therefore, the commodity relevant information is not real-time corresponding in the Query Result that sends of the commodity classification server of rear end and the current front-end server.For example: above-mentioned fixed cycle operator can be for once a day, and for 8:00 morning of every day operates accordingly, the user then inquires about at late 20:00.At this moment, the data of the commodity classification querying server institute foundation of rear end are the data that the user inquires about 8:00 morning on the same day.If the morning of this day 8:00 between the late 20:00, the relevant information of front-end merchandise changes, for example: commodity ID changes, and makes cabinet etc. under a certain commodity class commodity now owing to substituting season, and then the user can't correctly find and want the commodity inquired about.
So front-end server need screen the Query Result that receives, the commodity classification information obtained and current all commodity classifications that meet this query word are compared.
When the commodity classification that obtains and current commodity classification information conforms, front-end server shows this commodity classification information, and the DISPLAY ORDER of commodity classification from left to right is arranged in order according to the weight order from high to low of commodity classification correspondence.
When the commodity classification that obtains and current commodity classification information did not meet, front-end server shielded this commodity classification information, is not shown.
The embodiment of the present application provides a kind of querying method of vertical search, as shown in Figure 8, specifically may further comprise the steps:
Herein, it is query word that our definition is used for carrying out the required Query Information of one query, also can comprise the commodity classification.
Front-end server provides with the user and carries out mutual window by query page.The user visits this query page by browser, and the Query Information that input is wanted to inquire about in this query page is inquired about.
For example, when the user need buy a T-shirt, the user logined the online transaction website, and by the query page of this website, the input user wants the Query Information " T-shirt " inquired about.At this moment, when if the user only wants to check the information of relevant men's clothing T-shirt, the combobox that system ejects in the time of can passing through input " T-shirt " is selected the commodity classification, thereby in the input inquiry word, input commodity classification limits query context, and for example: the user selects " the T-shirt men's clothing " in the combobox, wherein " men's clothing " is the commodity classification, and Query Information is " T-shirt men's clothing ".
The user is input inquiry word " T-shirt " only, when providing " men's clothing ", " women's dress " and " easy dress " on the page that system returns according to " T-shirt ", the user selects " men's clothing ", and wherein " men's clothing " is the commodity classification, and Query Information is " T-shirt men's clothing ".
Wherein, front-end server is transmitted to commodity classification querying server with the Query Information of user input and may further comprise the steps:
(1) front-end server obtains the Query Information of user's input by query page.
(2) front-end server obtains the corresponding commodity classification of the commodity classification ID in this Query Information.
(3) front-end server is transmitted to query word in the Query Information and commodity classification ID the commodity classification server of rear end.
Wherein, the inquiry of commodity classification querying server is to inquire about according to the statistical data analysis that a last time period obtains, and before this inquiry, may further comprise the steps:
(1) commodity classification querying server will be compiled into queries dictionary by the inquiry file that statistic analysis result is obtained.
(2) commodity classification querying server regularly loads queries dictionary.
Queries dictionary loads and to refer to when startup of server, and the mode by memory-mapped directly is mapped to internal memory to the queries dictionary file with the compiling of hash algorithm Memory Mapping File and its form.
After commodity classification querying server obtained this user's Query Information, according to the data that are used for inquiry, namely queries dictionary was inquired about.
After commodity are clicked, all will generate corresponding daily record according to this time click behavior, be kept in the log server.Wherein, click behavior each time is a click logs, and the form of daily record is as shown in table 1.Wherein, query word is the query word in the Query Information of user's input, the user clicks and to check the commodity classification and select commodity to click from click the page that this commodity classification returns and check, this moment, this commodity classification was the inquiry classification, and namely the user clicks and checks the commodity the last commodity classification of clicking before.Commodity ID is the corresponding ID of each commodity number, is used for each commodity of unique identification.Item property is the corresponding satellite informations of these commodity, for example: brand.
The generation of each bar daily record record is along with user's click checks that log server regularly should all daily records in the time period import to distributed document storage and concurrent operation platform and produced simultaneously.Wherein, the regular operation of log server can be every day, or per 12 hours etc.
Log server is clicked the analytic statistics program and will regularly be carried out statistical study to stored log after regularly the daily record that generates being imported to distributed document storage and concurrent operation platform, draws statistic analysis result.The object of this statistical study can be nearest ten days daily record, or the daily record in nearest two weeks, and the concrete time can be adjusted according to empirical value or statistical demand.
Click the analytic statistics program and be arranged in distributed document storage and concurrent operation platform, be used for stored log is carried out statistical study.Concrete, click the analytic statistics program timing stored log is carried out statistical study, draw statistic analysis result, specifically may further comprise the steps:
(1) obtains the daily record of carrying out statistical study.
Log server can regularly upgrade the daily record in distributed document storage and the concurrent operation platform, when click analytic statistics program timing is carried out statistical study to stored log, need obtain the daily record of renewal, to obtain up-to-date daily record, improves statistical accuracy.
(2) query word is carried out normalized.
Because the query word of each user's input not necessarily meets the statistical standard of clicking the analytic statistics program, for the ease of adding up according to query word, need carry out normalized to query word.Normalized comprises removes unnecessary word in the query word, and the conversion of alphabet size between writing carried out, the conversion between the full-shape half-angle, the conversion between the simplified traditional font, the conversion of punctuate and the conversion between the Chinese figure etc. in unnecessary space.Query word through normalized can be directly used in the statistics of clicking the analytic statistics program.
(3) data are carried out in daily record and gather, generate and click distributed data.
When thousands of user inquires about, can import identical query word and click identical inquiry classification, according to the daily record in a period of time, query word and inquiry classification that the user is inquired about gather, and obtain the number of times of the commodity classification of clicking according to each query word.
(4) obtain commodity classification tree.
(5) in conjunction with commodity classification tree, generate click classification tree according to clicking distributed data.
(6) generate the inquiry file corresponding with clicking the classification tree.
(7) click the analytic statistics program with inquiry file as a result of, export.The inquiry file of this output will send to the commodity classification querying server of rear end, so that the commodity classification querying server of rear end generates corresponding queries dictionary according to this inquiry file, be used for follow-up inquiry.
The embodiment of the present application provides a kind of inquiry system of vertical search, comprising:
Querying server is for the Query Information that obtains the user; Obtain Query Result according to described Query Information in the queries dictionary of described commodity classification server, described Query Result for the commodity class that is complementary with described Query Information now sub-commodity classification and corresponding weight; Described Query Result is sent to described user, described user is checked; Obtain the statistic analysis result that Analysis server sends, upgrade the queries dictionary of described commodity classification server according to described statistic analysis result, be used for follow-up inquiry;
Log server, the commodity classification and the described Query Information that are used for checking according to described user generate daily record, and described daily record is sent to described Analysis server.
Analysis server is used for receiving the described daily record that described log server sends; Statistical study is carried out in described daily record obtained described statistic analysis result; Described statistical study binded up one's hair give described querying server.
Wherein, described Query Information comprises query word and commodity classification, and log server 900 as shown in Figure 9, comprising:
Sending module 930, be used for that the described daily record that generation module 920 generates is sent to described Analysis server and carry out statistical study acquisition statistic analysis result, described statistic analysis result is used for upgrading the queries dictionary of described commodity classification server, is used for follow-up inquiry.
Wherein, Analysis server 1000 as shown in figure 10, comprising:
Sending module 1030, the described statistic analysis result that is used for statistical analysis module 1020 is obtained sends to querying server, makes described querying server upgrade the queries dictionary of described querying server, is used for follow-up inquiry.
Wherein, statistical analysis module 1020 as shown in figure 11, comprising:
Statistical study submodule 1021 is used for carrying out statistical study according to the described daily record that the described acquisition module in the Preset Time obtains, and obtains statistic analysis result, and described statistic analysis result is commodity classification and the corresponding weight that described user checks; Described weight comprises the number of clicks of the commodity classification correspondence that described user checks and corresponding click probability in commodity classification at the same level;
Generate submodule 1022: be used for according to commodity classification tree the described statistic analysis result generated query file that statistical study submodule 1021 is obtained.
The embodiment of the present application provides a kind of server 1200, as querying server, is applied to comprise in the system of querying server, Analysis server and log server, as shown in figure 12, comprising:
Sending module 1230, sub-commodity classification for the described Query Result that enquiry module 1220 is obtained sorts according to its corresponding weight, and ranking results sent to described user, described user is checked, and make log server generate daily record according to commodity classification and the described Query Information that described user checks, and described daily record is sent to Analysis server carry out statistical study and obtain statistic analysis result, described statistic analysis result is used for upgrading the queries dictionary of described commodity classification server, is used for follow-up inquiry;
Wherein, sending module 1230 also is used for: set a weight threshold, the sub-commodity classification of the weight in the described Query Result greater than described weight threshold sorted according to its corresponding weight.
Wherein, described Query Information comprises through the described query word of normalized and the commodity classification ID of described commodity classification correspondence;
As shown in figure 13, also comprise concatenation module 1250, be used for described Query Result is spliced that described splicing comprises the commodity classification that obtains the ID of commodity classification described in described Query Result correspondence.
The application has improved the correlativity of Query Result and user inquiry by click the Query Result that record returns the user according to the user.Certainly, arbitrary product of enforcement the application might not need to reach simultaneously above-described all advantages.
Through the above description of the embodiments, those skilled in the art can be well understood to the application and can realize by the mode that software adds essential general hardware platform, can certainly pass through hardware, but the former is better embodiment under a lot of situation.Based on such understanding, the part that the application's technical scheme contributes to prior art in essence in other words can embody with the form of software product, this computer software product is stored in the storage medium, comprise that some instructions are with so that a station terminal equipment (can be mobile phone, personal computer, server, the perhaps network equipment etc.) carry out the described method of each embodiment of the application.
The above only is the application's preferred implementation; should be pointed out that for those skilled in the art, under the prerequisite that does not break away from the application's principle; can also make some improvements and modifications, these improvements and modifications also should be looked the application's protection domain.
Claims (11)
1. the querying method of a vertical search is applied to comprise in the system of querying server, Analysis server and log server, front-end server and commodity classification server, it is characterized in that, comprising:
Front-end server obtains user's Query Information and is transmitted to querying server, and Query Information comprises user's query word and commodity classification; Front-end server carries out normalized to query word and obtains the commodity ID of commodity classification correspondence; Front-end server will be transmitted to commodity classification server through query word and the described commodity classification ID of normalized;
Querying server obtains the user's of described front-end server transmission Query Information;
Described querying server obtains Query Result according to described Query Information in the queries dictionary of described commodity classification server, described Query Result for the commodity class that is complementary with described Query Information now sub-commodity classification and corresponding weight;
Described querying server sorts the sub-commodity classification in the described Query Result according to its corresponding weight, and ranking results sent to described user, described user is checked, and make log server generate daily record according to commodity classification and the described Query Information that described user checks, and described daily record is sent to Analysis server carry out statistical study and obtain statistic analysis result, described statistic analysis result is used for upgrading the queries dictionary of described commodity classification server, is used for follow-up inquiry.
2. the method for claim 1, it is characterized in that, at described querying server before corresponding weight sorts according to it with the sub-commodity classification in the described Query Result, also comprise: set a weight threshold, the sub-commodity classification of the weight in the described Query Result greater than described weight threshold sorted according to its corresponding weight.
3. the method for claim 1 is characterized in that,
The commodity classification that described log server is checked according to described user and described Query Information generate before the daily record, also comprise:
Obtain the described user's of front-end server forwarding Query Information;
Analysis server carries out statistical study and obtains statistic analysis result, and described statistic analysis result is used for upgrading the queries dictionary of described commodity classification server, is used for follow-up inquiry and specifically comprises:
Receive the interior daily record of Preset Time of described log server timed sending;
Carry out statistical study according to the described daily record in the Preset Time, obtain statistic analysis result, described statistic analysis result is commodity classification and the corresponding weight that described user checks; Described weight comprises the number of clicks of the commodity classification correspondence that described user checks and corresponding click probability in commodity classification at the same level;
According to commodity classification tree, with described statistic analysis result generated query file;
Described inquiry file is sent to described querying server, so that described querying server upgrades the queries dictionary of described commodity classification server, the inquiry that the user is follow-up according to described inquiry file.
4. the method for claim 1 is characterized in that, described querying server sorts the sub-commodity classification in the described Query Result according to its corresponding weight, and ranking results is sent to before the described user, also comprises:
Described querying server splices described Query Result, and described splicing comprises the commodity classification that obtains the ID of commodity classification described in described Query Result correspondence.
5. the inquiry system of a vertical search is characterized in that, comprising:
Querying server is for the Query Information that obtains the user; Obtain Query Result according to described Query Information in the queries dictionary of described commodity classification server, described Query Result for the commodity class that is complementary with described Query Information now sub-commodity classification and corresponding weight; Sub-commodity classification in the described Query Result is sorted according to its corresponding weight, and ranking results is sent to described user, described user is checked; Obtain the statistic analysis result that Analysis server sends, upgrade the queries dictionary of described commodity classification server according to described statistic analysis result, be used for follow-up inquiry;
Log server, the commodity classification and the described Query Information that are used for checking according to described user generate daily record, and described daily record is sent to described Analysis server.
Analysis server is used for receiving the described daily record that described log server sends; Statistical study is carried out in described daily record obtained described statistic analysis result; Described statistical study binded up one's hair give described querying server;
Front-end server is used for, and obtains at described querying server before user's the Query Information, obtains user's Query Information and is transmitted to described querying server, and described Query Information comprises described user's query word and commodity classification; Described front-end server carries out normalized to described query word and obtains the commodity ID of commodity classification correspondence;
Commodity classification server is for described query word and the described commodity classification ID of the process normalized that receives described front-end server transmission.
6. system as claimed in claim 5 is characterized in that, described Query Information comprises query word and commodity classification, and described log server comprises:
Acquisition module is for the user's who obtains the front-end server forwarding Query Information;
Generation module, the commodity classification and the described Query Information that are used for checking according to described user generate daily record;
Sending module is used for that the described daily record that described generation module generates is sent to described Analysis server and carries out statistical study acquisition statistic analysis result, and described statistic analysis result is used for upgrading the queries dictionary of described commodity classification server, is used for follow-up inquiry.
7. system as claimed in claim 5 is characterized in that, described Analysis server comprises:
Receiver module is used for receiving the daily record that described log server sends;
Statistical analysis module is used for statistical study is carried out in the described daily record that described receiver module receives, and obtains statistic analysis result;
Sending module, the described statistic analysis result that is used for described statistical analysis module is obtained sends to querying server, makes described querying server upgrade the queries dictionary of described querying server, is used for follow-up inquiry.
8. system as claimed in claim 7 is characterized in that, described statistical analysis module comprises:
The statistical study submodule is used for carrying out statistical study according to the described daily record that the described acquisition module in the Preset Time obtains, and obtains statistic analysis result, and described statistic analysis result is commodity classification and the corresponding weight that described user checks; Described weight comprises the number of clicks of the commodity classification correspondence that described user checks and corresponding click probability in commodity classification at the same level;
Generate submodule: be used for according to commodity classification tree the described statistic analysis result generated query file that described statistical study submodule is obtained.
9. a server as querying server, is applied to comprise in the system of querying server, Analysis server and log server, front-end server and commodity classification server, it is characterized in that, comprising:
Acquisition module is for the user's who obtains described front-end server forwarding Query Information; Be used for obtaining the statistic analysis result that Analysis server sends; Described acquisition module obtains before user's the Query Information, also comprises: front-end server obtains user's Query Information, and described Query Information comprises described user's query word and commodity classification; Described front-end server carries out normalized to described query word and obtains the commodity ID of commodity classification correspondence; Described front-end server will be transmitted to described commodity classification server through described query word and the described commodity classification ID of normalized;
Enquiry module, the described Query Information that is used for obtaining according to described acquisition module obtains Query Result in the queries dictionary of described commodity classification server, described Query Result for the commodity class that is complementary with described Query Information now sub-commodity classification and corresponding weight;
Sending module, sub-commodity classification for the described Query Result that described enquiry module is obtained sorts according to its corresponding weight, and ranking results sent to described user, described user is checked, and make log server generate daily record according to commodity classification and the described Query Information that described user checks, and described daily record is sent to Analysis server carry out statistical study and obtain statistic analysis result, described statistic analysis result is used for upgrading the queries dictionary of described commodity classification server, is used for follow-up inquiry;
Update module is used for upgrading according to the statistic analysis result that described acquisition module obtains the queries dictionary of described commodity classification server, and the queries dictionary after the described renewal is sent to described enquiry module, is used for follow-up inquiry.
10. server as claimed in claim 9 is characterized in that, described sending module also is used for: set a weight threshold, the sub-commodity classification of the weight in the described Query Result greater than described weight threshold sorted according to its corresponding weight.
11. server as claimed in claim 9 is characterized in that,
Described Query Information comprises the commodity classification ID through the described query word of normalized and described commodity classification correspondence;
Also comprise concatenation module, be used for described Query Result is spliced that described splicing comprises the commodity classification that obtains the ID of commodity classification described in described Query Result correspondence.
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 200910210422 CN102053983B (en) | 2009-11-02 | 2009-11-02 | Method, system and device for querying vertical search |
| HK11109117.6A HK1154967A1 (en) | 2009-11-02 | 2011-08-29 | A query method, system and device for verticle search |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 200910210422 CN102053983B (en) | 2009-11-02 | 2009-11-02 | Method, system and device for querying vertical search |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN102053983A CN102053983A (en) | 2011-05-11 |
| CN102053983B true CN102053983B (en) | 2013-09-25 |
Family
ID=43958320
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN 200910210422 Active CN102053983B (en) | 2009-11-02 | 2009-11-02 | Method, system and device for querying vertical search |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN102053983B (en) |
| HK (1) | HK1154967A1 (en) |
Families Citing this family (26)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103034665B (en) * | 2011-10-10 | 2016-01-06 | 阿里巴巴集团控股有限公司 | Information query method and device |
| CN102375885A (en) * | 2011-10-21 | 2012-03-14 | 北京百度网讯科技有限公司 | Method and device for providing search suggestions corresponding to query sequence |
| CN103176995B (en) * | 2011-12-21 | 2016-04-06 | 阿里巴巴集团控股有限公司 | A kind of method of information navigation, equipment and system |
| CN103577413B (en) * | 2012-07-20 | 2017-11-17 | 阿里巴巴集团控股有限公司 | Search result ordering method and system, search results ranking optimization method and system |
| CN102841946B (en) * | 2012-08-24 | 2016-05-25 | 北京国政通科技有限公司 | Commodity data retrieval ordering and Method of Commodity Recommendation and system |
| CN103678365B (en) * | 2012-09-13 | 2017-07-18 | 阿里巴巴集团控股有限公司 | The dynamic acquisition method of data, apparatus and system |
| CN103729362B (en) * | 2012-10-12 | 2017-07-21 | 阿里巴巴集团控股有限公司 | The determination method and apparatus of navigation content |
| CN104239021B (en) * | 2013-06-21 | 2017-12-08 | 阿里巴巴集团控股有限公司 | The generation method and device and search engine system of search engine inquiry string |
| CN104424296B (en) * | 2013-09-02 | 2018-07-31 | 阿里巴巴集团控股有限公司 | Query word sorting technique and device |
| CN103500231A (en) * | 2013-10-25 | 2014-01-08 | 乐视网信息技术（北京）股份有限公司 | Multi-media file recommending method and device |
| CN103530408A (en) * | 2013-10-25 | 2014-01-22 | 乐视网信息技术（北京）股份有限公司 | Multimedia file recommendation method and device |
| CN103678092B (en) * | 2013-12-30 | 2017-01-25 | 北京网康科技有限公司 | log analysis method and system |
| CN104462554B (en) * | 2014-12-25 | 2019-03-08 | 北京奇虎科技有限公司 | Question and answer page relevant issues recommended method and device |
| CN104462553B (en) * | 2014-12-25 | 2019-02-26 | 北京奇虎科技有限公司 | Question and answer page relevant issues recommended method and device |
| CN104462556B (en) * | 2014-12-25 | 2018-02-23 | 北京奇虎科技有限公司 | Question and answer page relevant issues recommend method and apparatus |
| CN104657733B (en) * | 2015-02-14 | 2016-04-06 | 冯贵良 | A kind of massive medical image data memory storage and method |
| CN105468729A (en) * | 2015-11-23 | 2016-04-06 | 深圳大粤网络视界有限公司 | Internet mobile vertical search engine |
| CN105468782B (en) * | 2015-12-21 | 2019-05-17 | 北京奇虎科技有限公司 | A kind of method and device of the resource matched degree judgement of inquiry- |
| CN106933814A (en) * | 2015-12-28 | 2017-07-07 | 航天信息股份有限公司 | Tax data exception analysis method and system |
| CN108205726A (en) * | 2016-12-19 | 2018-06-26 | 北京京东尚科信息技术有限公司 | Data analysis processing method, system and device |
| WO2018165878A1 (en) * | 2017-03-14 | 2018-09-20 | 深圳市瑞荣创电子科技有限公司 | Device for providing commodity supply information in business circle |
| CN107066544A (en) * | 2017-03-14 | 2017-08-18 | 深圳市瑞荣创电子科技有限公司 | A kind of business circles' supply of commodities information provider unit |
| CN112669122A (en) * | 2017-05-11 | 2021-04-16 | 微重力（北京）科技有限公司 | Article query comparison method and device, storage medium and processor |
| CN108256719B (en) * | 2017-07-24 | 2022-11-04 | 平安科技（深圳）有限公司 | Resource processing system and method |
| CN108510321A (en) * | 2018-03-23 | 2018-09-07 | 北京焦点新干线信息技术有限公司 | A kind of construction method and device of house property user portrait |
| CN110110044B (en) * | 2019-04-11 | 2020-05-05 | 广州探迹科技有限公司 | Method for enterprise information combination screening |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101079071A (en) * | 2007-06-20 | 2007-11-28 | 华为技术有限公司 | Advertisement correlation method of vertical search engine and vertical search advertisement system |
| CN101409748A (en) * | 2008-07-08 | 2009-04-15 | 浙江大学 | System and method for collecting, indexing, subscribing and publishing mobile terminal information |
-
2009
- 2009-11-02 CN CN 200910210422 patent/CN102053983B/en active Active
-
2011
- 2011-08-29 HK HK11109117.6A patent/HK1154967A1/en unknown
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101079071A (en) * | 2007-06-20 | 2007-11-28 | 华为技术有限公司 | Advertisement correlation method of vertical search engine and vertical search advertisement system |
| CN101409748A (en) * | 2008-07-08 | 2009-04-15 | 浙江大学 | System and method for collecting, indexing, subscribing and publishing mobile terminal information |
Also Published As
| Publication number | Publication date |
|---|---|
| CN102053983A (en) | 2011-05-11 |
| HK1154967A1 (en) | 2012-05-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102053983B (en) | Method, system and device for querying vertical search | |
| US10789311B2 (en) | Method and device for selecting data content to be pushed to terminal, and non-transitory computer storage medium | |
| CN102236663B (en) | Query method, query system and query device based on vertical search | |
| JP5721818B2 (en) | Use of model information group in search | |
| CN104850546B (en) | Display method and system of mobile media information | |
| US9798820B1 (en) | Classification of keywords | |
| CN104102639B (en) | Popularization triggering method based on text classification and device | |
| CN111008265A (en) | Enterprise information searching method and device | |
| CN104216881A (en) | Method and device for recommending individual labels | |
| CN103838756A (en) | Method and device for determining pushed information | |
| CN103034680B (en) | For data interactive method and the device of terminal device | |
| CN105279224A (en) | Information push method and device | |
| CN104077286A (en) | Commodity information search method and system | |
| KR20140026932A (en) | System and method providing a suited shopping information by analyzing the propensity of an user | |
| CN103118111A (en) | Information push method based on data from a plurality of data interaction centers | |
| CN1983255A (en) | Internet searching method | |
| CN104077407A (en) | System and method for intelligent data searching | |
| US11663280B2 (en) | Search engine using joint learning for multi-label classification | |
| CN103020128B (en) | With the method and apparatus of data interaction with terminal device | |
| CN112818230B (en) | Content recommendation method, device, electronic equipment and storage medium | |
| CN114881712A (en) | Intelligent advertisement putting method, device, equipment and storage medium | |
| CN108470289B (en) | Virtual article issuing method and equipment based on E-commerce shopping platform | |
| CN105159898A (en) | Searching method and searching device | |
| CN116823410B (en) | Data processing method, object processing method, recommending method and computing device | |
| CN116308683B (en) | Knowledge-graph-based clothing brand positioning recommendation method, equipment and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| REG | Reference to a national code |
Ref country code: HK Ref legal event code: DE Ref document number: 1154967 Country of ref document: HK |
|
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| REG | Reference to a national code |
Ref country code: HK Ref legal event code: GR Ref document number: 1154967 Country of ref document: HK |